eazygrad.randn¶

eazygrad.randn(*shape: int, requires_grad: bool = False, dtype: Any = numpy.float32) → _Tensor[source]¶

Return a tensor filled with samples from the standard normal distribution.

Parameters:

*shape (int) – Output shape.
requires_grad (bool, default=False) – Whether to track gradients for the returned tensor.
dtype (numpy.dtype or type, default=numpy.float32) – Output dtype.

Returns:

_Tensor – Randomly initialized tensor.